With the rapid development of the science and technology, computer vision and/or image processing technologies have been gradually used in positioning of various electronic products (e.g., various mobile apparatus, and wearable equipments of augmented reality) and, thereby, providing various services for users depending on the positions of the users (e.g., positioning the location of the user and providing a guided path for the user when he/she is visiting an exhibit).
The conventional visual positioning technology combines the special image label (e.g., a visual code marker) with the image detecting technology. Specifically, the conventional technology uses an electronic product with the camera function to sense a visual code marker to derive an image. Then, the electronic product identifies a marker image comprised in the image and calculates the position of the electronic product in the space by the computer visual technology according to four corner positions of the marker image, the actual size of the visual code marker, and the camera calibration information. However, the image of the visual code marker captured by the electronic product may be vague, which results in the situation that the marker image cannot be identified accurately and/or the four corner positions of the marker image cannot be detected accurately. In this case, the electronic product cannot be positioned accurately.
Accordingly, the technique that can reduce the inaccuracy of a sensed image of a special image label to obtain a more accurate positioning result is still in an urgent need.